Skip to content

01Image verification

One check for AI generation, manipulation and provenance.

Fileverity runs several independent analyses on every image and combines them into one conclusion your team can act on, with the evidence attached.

01Inspection

Look closer, down to the pixel.

Move the loupe across the photo. It magnifies the pixels under the crosshair and measures their position, color and local variation. R1 marks a patch of rust; R2 marks its copy on another panel.

8×
X
—
Y
—
RGB
—
HEX
—
σ 15×15
—
Region
—

Demonstration image. We edited this photo ourselves, copying a patch of rust onto another panel (a copy-move manipulation).

02

What every analysis covers

  • 01 / 06

    AI image detection

    Several independent pixel-level detectors estimate whether the whole image was created or re-rendered by a generative model. Scores are calibrated, so they mean the same thing from one image to the next.

  • 02 / 06

    Generator attribution

    When the evidence allows it, the report names the likely model family and the source of that attribution, such as metadata, a watermark or pixel analysis.

  • 03 / 06

    Manipulation localization

    Independent forensic analyses look for splicing, copy-move, retouching and inpainting, then mark suspicious regions on the image.

  • 04 / 06

    Editing-software traces

    Reads file-level records left by tools such as Photoshop, Lightroom and GIMP, and separates content changes from files that were only re-saved.

  • 05 / 06

    Provenance and AI labels

    Reads Content Credentials (C2PA), EXIF and XMP metadata, generation parameters and machine-readable AI labels, such as those required under China's GB 45438, then checks them against the pixels.

  • 06 / 06

    Evidence fusion

    All signals are weighed together. Contradictions are surfaced, not hidden, and the result can be inconclusive when the evidence is insufficient.

03

A clear answer, not a raw score

Each image receives one of four conclusions, an authenticity score from 0 to 100 and a risk level that maps to a recommended action under your chosen policy.

  • 01Authentic, no manipulation found
  • 02Authentic with local manipulation
  • 03AI-generated
  • 04Inconclusive

04

Two ways to verify

01

Quick Check

A fast online check for a single image: an AI-generation verdict, an editing flag and a heatmap preview of suspicious areas.

  • Results in seconds
  • Free account: three checks and one Pro analysis
  • Upload, drag and drop or paste

02

Pro analysis

The full evidence pipeline: generator attribution, region masks, verifiable reports, batch processing and API access.

  • All detectors and forensic checks
  • Verifiable PDF and JSON reports
  • API, batch and review workflow

05

Fits into existing workflows

Use the web app for case-by-case review, or call the REST API from your claims, onboarding or moderation system. Reports come as JSON for systems and as print-ready PDFs for people.

  • REST API with JSON responses
  • Report verification endpoint
  • Case reference fields for your own IDs
  • Reviewer decisions stored as revisions

Also available: invisible watermarking

Embed an invisible, keyed watermark in the images you publish, then detect it later to prove origin and trace redistribution.

06

Frequently asked questions

Q01

What types of image manipulation can Fileverity detect?

Fileverity detects splicing, copy-move, object removal, retouching, text changes and AI inpainting. It also reports the file-level editing history recorded by editing software.

Q02

How do AI-generated and manipulated images differ?

An AI-generated image was produced or re-rendered as a whole by a generative model, while a manipulated image is a real photo that was changed in specific regions. Fileverity reports the two separately, so reviewers know which question each finding answers.

Q03

Can Fileverity verify C2PA Content Credentials?

Yes. Fileverity reads and validates C2PA Content Credentials and other embedded provenance data, then checks whether the claims in the file are consistent with the pixel analysis.

Q04

Does Fileverity have an API?

Yes. Pro and Enterprise plans include a REST API that returns structured results and reports, plus an endpoint for verifying any report later.

Q05

Can Fileverity detect deepfakes?

Yes, for still images. Fileverity detects images that were generated or re-rendered by AI and locates AI-edited regions, which covers most image deepfakes. Dedicated face-swap detection is in development.

Put evidence behind every image decision.

Tell us about your workflow and volume. We will set up access and walk your team through real cases.